Github user pnowojski commented on a diff in the pull request:
https://github.com/apache/flink/pull/4239#discussion_r134673133
--- Diff:
flink-connectors/flink-connector-kafka-0.11/src/main/java/org/apache/flink/streaming/connectors/kafka/FlinkKafkaProducer011.java
---
@@ -0,0 +1,1000 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.streaming.connectors.kafka;
+
+import org.apache.flink.api.common.functions.RuntimeContext;
+import org.apache.flink.api.common.state.ListState;
+import org.apache.flink.api.common.state.ListStateDescriptor;
+import org.apache.flink.api.common.typeinfo.TypeHint;
+import org.apache.flink.api.common.typeinfo.TypeInformation;
+import org.apache.flink.api.java.ClosureCleaner;
+import org.apache.flink.api.java.typeutils.GenericTypeInfo;
+import org.apache.flink.configuration.Configuration;
+import org.apache.flink.metrics.MetricGroup;
+import org.apache.flink.runtime.state.FunctionInitializationContext;
+import org.apache.flink.runtime.state.FunctionSnapshotContext;
+import org.apache.flink.runtime.util.SerializableObject;
+import org.apache.flink.streaming.api.datastream.DataStream;
+import org.apache.flink.streaming.api.datastream.DataStreamSink;
+import
org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator;
+import
org.apache.flink.streaming.api.functions.sink.TwoPhaseCommitSinkFunction;
+import org.apache.flink.streaming.api.operators.StreamSink;
+import org.apache.flink.streaming.api.operators.StreamingRuntimeContext;
+import
org.apache.flink.streaming.connectors.kafka.internal.FlinkKafkaProducer;
+import
org.apache.flink.streaming.connectors.kafka.internal.metrics.KafkaMetricMuttableWrapper;
+import
org.apache.flink.streaming.connectors.kafka.partitioner.FlinkFixedPartitioner;
+import
org.apache.flink.streaming.connectors.kafka.partitioner.FlinkKafkaDelegatePartitioner;
+import
org.apache.flink.streaming.connectors.kafka.partitioner.FlinkKafkaPartitioner;
+import org.apache.flink.streaming.runtime.streamrecord.StreamRecord;
+import
org.apache.flink.streaming.util.serialization.KeyedSerializationSchema;
+import
org.apache.flink.streaming.util.serialization.KeyedSerializationSchemaWrapper;
+import org.apache.flink.streaming.util.serialization.SerializationSchema;
+import org.apache.flink.util.ExceptionUtils;
+import org.apache.flink.util.NetUtils;
+
+import org.apache.kafka.clients.producer.Callback;
+import org.apache.kafka.clients.producer.Producer;
+import org.apache.kafka.clients.producer.ProducerConfig;
+import org.apache.kafka.clients.producer.ProducerRecord;
+import org.apache.kafka.clients.producer.RecordMetadata;
+import org.apache.kafka.common.Metric;
+import org.apache.kafka.common.MetricName;
+import org.apache.kafka.common.PartitionInfo;
+import org.apache.kafka.common.errors.InvalidTxnStateException;
+import org.apache.kafka.common.serialization.ByteArraySerializer;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.annotation.Nullable;
+
+import java.io.Closeable;
+import java.io.IOException;
+import java.io.Serializable;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.Comparator;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Map;
+import java.util.Properties;
+import java.util.Set;
+import java.util.UUID;
+import java.util.concurrent.BlockingDeque;
+import java.util.concurrent.LinkedBlockingDeque;
+import java.util.concurrent.atomic.AtomicLong;
+
+import static java.util.Objects.requireNonNull;
+
+/**
+ * Flink Sink to produce data into a Kafka topic. This producer is
compatible with Kafka 0.11.x. By default producer
+ * will use {@link Semantic#EXACTLY_ONCE} semantic.
+ *
+ * <p>Implementation note: This producer is a hybrid between a regular
regular
+ * {@link org.apache.flink.streaming.api.functions.sink.SinkFunction} (a)
and a custom operator (b).
+ *
+ * <p>Details about approach (a):
+ * Because of regular {@link
org.apache.flink.streaming.api.functions.sink.SinkFunction} APIs limitations,
this
+ * variant do not allow accessing the timestamp attached to the record.
+ *
+ * <p>Details about approach (b):
+ * Kafka 0.11 supports writing the timestamp attached to a record to
Kafka. When using the
+ * {@link FlinkKafkaProducer011#writeToKafkaWithTimestamps} method, the
Kafka producer can access the internal
+ * record timestamp of the record and write it to Kafka.
+ *
+ * <p>All methods and constructors in this class are marked with the
approach they are needed for.
+ */
+public class FlinkKafkaProducer011<IN>
+ extends TwoPhaseCommitSinkFunction<IN,
FlinkKafkaProducer011.KafkaTransactionState> {
+
+ /**
+ * Semantics that can be chosen.
+ * <li>{@link #EXACTLY_ONCE}</li>
+ * <li>{@link #AT_LEAST_ONCE}</li>
+ * <li>{@link #NONE}</li>
+ */
+ public enum Semantic {
+ /**
+ * Semantic.EXACTLY_ONCE the Flink producer will write all
messages in a Kafka transaction that will be
+ * committed to the Kafka on a checkpoint.
+ *
+ * <p>In this mode {@link FlinkKafkaProducer011} sets up a pool
of {@link FlinkKafkaProducer}. Between each
+ * checkpoint there is created new Kafka transaction, which is
being committed on
+ * {@link
FlinkKafkaProducer011#notifyCheckpointComplete(long)}. If checkpoint complete
notifications are
+ * running late, {@link FlinkKafkaProducer011} can run out of
{@link FlinkKafkaProducer}s in the pool. In that
+ * case any subsequent {@link
FlinkKafkaProducer011#snapshotState(FunctionSnapshotContext)} requests will fail
+ * and {@link FlinkKafkaProducer011} will keep using the {@link
FlinkKafkaProducer} from previous checkpoint.
+ * To decrease chances of failing checkpoints there are three
options:
+ * <li>decrease number of max concurrent checkpoints</li>
+ * <li>make checkpoints more reliable (so that they complete
faster)</li>
+ * <li>increase delay between checkpoints</li>
+ * <li>increase size of {@link FlinkKafkaProducer}s pool</li>
+ */
+ EXACTLY_ONCE,
+ /**
+ * Semantic.AT_LEAST_ONCE the Flink producer will wait for all
outstanding messages in the Kafka buffers
+ * to be acknowledged by the Kafka producer on a checkpoint.
+ */
+ AT_LEAST_ONCE,
+ /**
+ * Semantic.NONE means that nothing will be guaranteed.
Messages can be lost and/or duplicated in case
+ * of failure.
+ */
+ NONE
+ }
+
+ private static final Logger LOG =
LoggerFactory.getLogger(FlinkKafkaProducerBase.class);
+
+ private static final long serialVersionUID = 1L;
+
+ /**
+ * Default number of KafkaProducers in the pool. See {@link
Semantic#EXACTLY_ONCE}.
+ */
+ public static final int DEFAULT_KAFKA_PRODUCERS_POOL_SIZE = 5;
--- End diff --
As I remember the reason was that it is not easy/not possible at the moment
to get this information in the operator. It should be a follow up work.
Regardless of this, code of this operator would look the same (because we don't
have guarantees for the `notifyCheckpointComplete` to always reach us on time).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
